Interpreting Epidemiologic Evidence and Causal Inference in Obesity Research

Abstract This chapter discusses key conceptual issues related to interpretation of epidemiologic evidence and inferences of causation in obesity research. It begins by discussing the role of randomized controlled trials (RCTs). It then addresses several methodological issues in establishing causality, such as confounding, reverse causation, measurement errors, mediation and effect modification, validity versus generalizability, and the calculation and interpretation of population attributable risk (PAR). Finally, the chapter reviews commonly used criteria of causality in obesity epidemiology.